The Autodidactic Universe
arXiv:2104.03902
Abstract
We present an approach to cosmology in which the Universe learns its own physical laws. It does so by exploring a landscape of possible laws, which we express as a certain class of matrix models. We discover maps that put each of these matrix models in correspondence with both a gauge/gravity theory and a mathematical model of a learning machine, such as a deep recurrent, cyclic neural network. This establishes a correspondence between each solution of the physical theory and a run of a neural network. This correspondence is not an equivalence, partly because gauge theories emerge from limits of the matrix models, whereas the same limits of the neural networks used here are not well-defined. We discuss in detail what it means to say that learning takes place in autodidactic systems, where there is no supervision. We propose that if the neural network model can be said to learn without supervision, the same can be said for the corresponding physical theory. We consider other protocols for autodidactic physical systems, such as optimization of graph variety, subset-replication using self-attention and look-ahead, geometrogenesis guided by reinforcement learning, structural learning using renormalization group techniques, and extensions. These protocols together provide a number of directions in which to explore the origin of physical laws based on putting machine learning architectures in correspondence with physical theories.
arXiv comment79 pages, 11 figures
References in corpus (36)
- Relational Quantum Mechanics
- A class of quantum many-body states that can be efficiently simulated
- Density estimation using Real NVP
- The Spin Foam Approach to Quantum Gravity
- Hamiltonian Neural Networks
- Measuring thermodynamic length
- Learning Structural Node Embeddings Via Diffusion Wavelets
- M(atrix) Theory: Matrix Quantum Mechanics as a Fundamental Theory
- Mutual Information, Neural Networks and the Renormalization Group
- An exact mapping between the Variational Renormalization Group and Deep Learning
- Neural Network Renormalization Group
- Deep Learning and AdS/CFT
- Lagrangian Neural Networks
- Super-resolving the Ising model with convolutional neural networks
- Energetic Causal Sets
- AdS/CFT as a deep Boltzmann machine
- The Objective Past of a Quantum Universe: Redundant Records of Consistent Histories
- The Plebanski action extended to a unification of gravity and Yang-Mills theory
- Network Geometry and Complexity
- Operational General Relativity: Possibilistic, Probabilistic, and Quantum
- Deep learning and the renormalization group
- Ollivier-Ricci curvature convergence in random geometric graphs
- Is Deep Learning a Renormalization Group Flow?
- The exceptional Jordan algebra and the matrix string
- Sum of squares of degrees in a graph
- Numerical verification of the microscopic time reversibility of Newton's equations of motion: Fighting exponential divergence
- On The Construction of Zero Energy States in Supersymmetric Matrix Models III
- Emergence of the Circle in a Statistical Model of Random Cubic Graphs
- Extremal variety as the foundation of a cosmological quantum theory
- A Correspondence Between Random Neural Networks and Statistical Field Theory
- The significance of numerical coincidences in nature
- Convergence of Combinatorial Gravity
- BRST quantization of Matrix Chern-Simons Theory
- Precedence and freedom in quantum physics
- Unification of the state with the dynamical law
- Learning the Einstein-Podolsky-Rosen correlations on a Restricted Boltzmann Machine